Dataset Viewer
Auto-converted to Parquet Duplicate
word
stringlengths
0
4
start
float64
0
1.59k
end
float64
0.18
1.59k
0
0.5
0.5
0.72
0.72
1.16
1.16
1.4
1.4
1.4
1.4
1.64
1.64
2.1
2.1
2.1
2.36
2.88
2.88
2.98
2.98
3.24
先生
3.42
3.46
3.46
3.72
朋友
3.92
3.98
3.98
4.56
大家好
4.56
5.06
6.24
6.76
6.76
7.02
7.02
7.3
7.3
7.64
7.64
7.78
世界
7.78
8.44
8.44
8.94
8.94
8.94
8.94
9.22
9.22
9.4
9.4
10.52
10.52
10.52
10.52
10.84
10.84
11.1
12.6
13.12
13.12
13.22
13.22
13.46
13.46
13.84
中国
14.08
14.6
14.6
15
15
15.14
15.14
15.44
15.44
15.66
15.66
16
16
16.16
16.16
16.780001
16.780001
16.98
16.98
17.02
17.02
17.32
18.4
18.92
中国
18.92
19.24
19.24
19.5
19.5
19.68
19.68
20.200001
20.940001
21.1
21.1
21.459999
21.459999
21.639999
21.639999
21.74
21.74
22.02
22.02
22.18
22.18
22.26
22.26
22.700001
22.700001
22.700001
22.700001
23.32
中国
23.780001
24.299999
24.299999
24.540001
24.540001
24.719999
24.719999
24.98
24.98
25.299999
25.299999
25.440001
25.440001
25.639999
25.639999
25.860001
26.16
26.26
26.26
26.360001
26.360001
26.76
26.76
26.84
28.18
28.74
28.74
29.16
29.16
29.48
29.48
29.620001
29.620001
29.84
29.84
29.940001
29.940001
30.16
30.16
30.32
30.32
30.82
30.82
31.02
31.02
31.02
31.02
31.440001
31.440001
31.540001
我们
32.400002
32.959999
32.959999
33.400002
33.400002
33.560001
33.560001
33.68
33.68
34
34
34.18
34.720001
35.279999
35.279999
35.52
35.52
35.720001
35.720001
35.939999
36.599998
37.16
37.16
37.5
37.5
38.200001
38.200001
38.32
38.32
38.98
End of preview. Expand in Data Studio

习近平音源

所有声音资料来自公开影像,属于公有领域
目前有 1h30m 的截取后声音,足够进行 Fine-tuning

Usage

按句截取

python -m pip install -r requirement.txt
python split.py

新增声音资料后,使用 Whisper 产生带有时间标记的 JSON 档,并手动复制到 ./voice/[FILE].json

export OPENAI_API_KEY="API_KEY_HERE"
python whisper.py ./[FILE].[AUDIO_EXTENSION]

产生 Bert-VITS2 微调所需的 esd.list 档案

python index_to_list.py
Downloads last month
45

Models trained or fine-tuned on KritiAI/Xijinping-TTS-Voicebank